118 



BRIDGMAN. 



inch the product is Hkely to be very brittle, or break spontaneously 

 into small pieces. If the temperature is raised very close to the melt- 

 ing point, however, it is possible to get by extrusion a uniform straight 

 wire with no apparent flaws, and not as brittle as the casting. 



The thermo-couple holes were drilled in the four pieces in the 

 regular way. The heating elements, as in the case of bismuth, were 

 mounted directly in small holes drilled in the ends, it not being feasible 

 to drill so large a hole as the use of the copper capsule would have 

 demanded. This manner of attaching the heating element was 

 responsible for the greater irregularity of the points, as also in the case 

 of bismuth. In one case the scattering was such and the accidental 







I 2 3 4 3 7 ■-) 



Pressure. Kg./Cm.'X 10' 



Antimony 



Figure 11. Antimony- Thermal conductivity in arbitrary units against 

 pressure in thousands of kg/cm-. Results obtained with a longitudinal flow 

 specimen. 



distril)ution such that a positive sign for the effect might have been 

 suspected. 



The readings obtained with the first of the cast specimens were the 

 most regular; these are reproduced in Figure 11. The thermo-couple 

 used with the second of the extruded specimens was therlo-constantan, 

 instead of copper-constantan. The readings with this were essentially 

 the same as with the others, thus again showing freedom from heat 

 leak along the wires of the couple. 



The thermal conducti\dty decreases with rising pressure; the two 

 cast specimens gave respectively —23.9 and —26.3%, and the two 

 extruded specimens —24.8 and —23.9%. The mean of all four is 



